Tornadoes have caused billions of dollars in damage and are one of the leading causes of weather-related deaths in the United States each year. Tornadoes are known to frequently form in the warm sector of extratropical cyclones (ETCs), yet relatively little research exists quantifying the climatological relationship between the two phenomena and their covariability. This work analyzes the climatology of F/EF1+ tornadoes relative to ETCs (ETCTORs) using historical databases of tornadoes and hourly ETC centers for 1980-2022. Most tornadoes (72\%) occurred broadly to the southeast of an ETC center within 2000km, with a median distance of approximately 500km. Of those tornadoes, 69\% occurred in outbreaks of 6+ tornadoes. The spatial and ETC-relative distributions are similar across all intensity levels. Through the seasonal cycle, tornadoes shift north and south along with ETCs and the jet stream, and they are much more strongly (weakly) associated with ETCs in the winter (76\%; summer 38\%) when the jet stream and ETCs are strongest (weakest). Finally, median tornado and tornado-producing ETC locations covary strongly with one another interannually. A long-term eastward trend in torn
Accurately reconstructing the 3D structure of tornadoes is critically important for understanding and preparing for this highly destructive weather phenomenon. While modern 3D scene reconstruction techniques, such as 3D Gaussian splatting (3DGS), could provide a valuable tool for reconstructing the 3D structure of tornados, at present we are critically lacking a controlled tornado dataset with which to develop and validate these tools. In this work we capture and release a novel multiview dataset of a small lab-based tornado. We demonstrate one can effectively reconstruct and visualize the 3D structure of this tornado using 3DGS.
We propose a statistical approach to tornadoes modeling for predicting and simulating occurrences of tornadoes and accumulated cost distributions over a time interval. This is achieved by modeling the tornadoes intensity, measured with the Fujita scale, as a stochastic process. Since the Fujita scale divides tornadoes intensity into six states, it is possible to model the tornadoes intensity by using Markov and semi-Markov models. We demonstrate that the semi-Markov approach is able to reproduce the duration effect that is detected in tornadoes occurrence. The superiority of the semi-Markov model as compared to the Markov chain model is also affirmed by means of a statistical test of hypothesis. As an application we compute the expected value and the variance of the costs generated by the tornadoes over a given time interval in a given area. he paper contributes to the literature by demonstrating that semi-Markov models represent an effective tool for physical analysis of tornadoes as well as for the estimation of the economic damages to human things.
Observations in the 171 AA channel of the Atmospheric Imaging Assembly of the space-borne Solar Dynamics Observatory show tornadoes-like features in the atmosphere of the Sun. These giant tornadoes appear as dark, elongated and apparently rotating structures in front of a brighter background. This phenomenon is thought to be produced by rotating magnetic field structures that extend throughout the atmosphere. We characterize giant tornadoes through a statistical analysis of properties like spatial distribution, lifetimes, and sizes. A total number of 201 giant tornadoes are detected in a period of 25 days, suggesting that on average about 30 events are present across the whole Sun at a time close to solar maximum. Most tornadoes appear in groups and seem to form the legs of prominences, thus serving as plasma sources/sinks. Additional Halpha observations with the Swedish 1-m Solar Telescope imply that giant tornadoes rotate as a structure although clearly exhibiting a thread-like structure. We observe tornado groups that grow prior to the eruption of the connected prominence. The rotation of the tornadoes may progressively twist the magnetic structure of the prominence until it bec
Solar magnetized "tornadoes", a phenomenon discovered in the solar atmosphere, appear as tornado-like structures in the corona but root in the photosphere. Like other solar phenomena, solar tornadoes are a feature of magnetized plasma and therefore differ distinctly from terrestrial tornadoes. Here we report the first analysis of solar "tornadoes" {Two papers which focused on different aspect of solar tornadoes were published in the Astrophysical Journal Letters (Li et al. 2012) and Nature (Wedemeyer-Böhm et al. 2012), respectively, during the revision of this Letter.}. A detailed case study of two events indicates that they are rotating vertical magnetic structures probably driven by underlying vortex flows in the photosphere. They usually exist as a group and relate to filaments/prominences, another important solar phenomenon whose formation and eruption are still mysteries. Solar tornadoes may play a distinct role in the supply of mass and twists to filaments. These findings could lead to a new explanation to filament formation and eruption.
We studied the dynamics of all prominence tornadoes detected by the Solar Dynamics Observatory/Atmospheric Imaging Assembly from 2011 January 01 to December 31. In total, 361 events were identified during the whole year, but only 166 tornadoes were traced until the end of their lifetime. Out of 166 tornadoes, 80 (48%) triggered CMEs in hosting prominences, 83 (50%) caused failed coronal mass ejections (CMEs) or strong internal motion in the prominences, and only 3 (2%) finished their lifetimes without any observed activity. Therefore, almost all prominence tornadoes lead to the destabilization of their hosting prominences and half of them trigger CMEs. Consequently, prominence tornadoes may be used as precursors for CMEs and hence for space weather predictions.
Developing methods to predict disastrous natural phenomena is more important than ever, and tornadoes are among the most dangerous ones in nature. Due to the unpredictability of the weather, counteracting them is not an easy task and today it is mainly carried out by expert meteorologists, who interpret meteorological models. In this paper we propose a system for the early detection of a tornado, validating its effectiveness in a real-world context and exploiting meteorological data collection systems that are already widespread throughout the world. Our system was able to predict tornadoes with a maximum probability of 84% up to five days before the event on a novel dataset of more than 5000 tornadic and non-tornadic events. The dataset and the code to reproduce our results are available at: https://tinyurl.com/3brsfwpk
The barbs or legs of some prominences show an apparent motion of rotation, which are often termed solar tornadoes. It is under debate whether the apparent motion is a real rotating motion, or caused by oscillations or counter-streaming flows. We present analysis results from spectroscopic observations of two tornadoes by the Interface Region Imaging Spectrograph. Each tornado was observed for more than 2.5 hours. Doppler velocities are derived through a single Gaussian fit to the Mg~{\sc{ii}}~k~2796Å~and Si~{\sc{iv}}~1393Å~line profiles. We find coherent and stable red and blue shifts adjacent to each other across the tornado axes, which appears to favor the interpretation of these tornadoes as rotating cool plasmas with temperatures of $10^4$ K-$10^5$ K. This interpretation is further supported by simultaneous observations of the Atmospheric Imaging Assembly on board the Solar Dynamics Observatory, which reveal periodic motions of dark structures in the tornadoes. Our results demonstrate that spectroscopic observations can provide key information to disentangle different physical processes in solar prominences.
Recent observations near the Galactic Centre have found several molecular filaments displaying striking helically-wound morphology, which are collectively known as "molecular tornadoes." We investigate the equilibrium structure of these molecular tornadoes by formulating a magnetohydrodynamic model of a rotating, helically magnetized filament. A special analytical solution is derived where centrifugal forces balance exactly with toroidal magnetic stress. From the physics of torsional Alfvén waves, we derive a constraint that links the toroidal flux-to-mass ratio and the pitch angle of the helical field to the rotation laws, which we find to be an important component in describing molecular tornado structure. The models are compared to the Ostriker solution for isothermal, non-magnetic, non-rotating filaments. We find that neither the analytic model nor the Alfvén wave model suffer from unphysical density inversions noted by other authors. A Monte Carlo exploration of our parameter space is constrained by observational measurements of the Pigtail Molecular Cloud (Pigtail), Double Helix Nebula (DHN), and Galactic Centre molecular Tornado (GCT). Observable properties such as the veloc
Based on an optical analogy of spintronics, the generation of optical tornadoes is theoretically investigated in two-dimensional photonic crystals without space-inversion symmetry. We address its close relation to the Berry curvature in crystal momentum space, which represents the non-trivial geometric property of a Bloch state. It is shown that the Berry curvature is easily controlled by tuning two types of dielectric rods in a honeycomb photonic crystal. Then, Bloch states with large Berry curvatures appear as optical tornadoes in real space. The radiation force of such a configuration is analyzed, and its possible application is also discussed.
Solar tornadoes are dark vertical filamentary structures observed in the extreme ultraviolet associated with prominence legs and filament barbs. Their true nature and relationship to prominences requires understanding their magnetic structure and dynamic properties. Recently, a controversy has arisen: is the magnetic field organized forming vertical, helical structures or is it dominantly horizontal? And concerning their dynamics, are tornadoes really rotating or is it just a visual illusion? Here, we analyze four consecutive spectropolarimetric scans of a prominence hosting tornadoes on its legs which help us shed some light on their magnetic and dynamical properties. We show that the magnetic field is very smooth in all the prominence, probably an intrinsic property of the coronal field. The prominence legs have vertical helical fields that show slow temporal variation probably related to the motion of the fibrils. Concerning the dynamics, we argue that 1) if rotation exists, it is intermittent, lasting no more than one hour, and 2) the observed velocity pattern is also consistent with an oscillatory velocity pattern (waves).
Tornadoes are the most violent of all atmospheric storms. In a typical year, the United States experiences hundreds of tornadoes with associated damages on the order of one billion dollars. Community preparation and resilience would benefit from accurate predictions of these economic losses, particularly as populations in tornado-prone areas increase in density and extent. Here, we use a zero-inflated modeling approach and artificial neural networks to predict tornado-induced property damage using publicly available data. We developed a neural network that predicts whether a tornado will cause property damage (out-of-sample accuracy = 0.821 and area under the receiver operating characteristic curve, AUROC, = 0.872). Conditional on a tornado causing damage, another neural network predicts the amount of damage (out-of-sample mean squared error = 0.0918 and R2 = 0.432). When used together, these two models function as a zero-inflated log-normal regression with hidden layers. From the best-performing models, we provide static and interactive gridded maps of monthly predicted probabilities of damage and property damages for the year 2019. Two primary weaknesses include (1) model fitting
In this paper we consider the role of nonmodal instabilities in the dynamics of atmospheric tornadoes. For this purpose we consider the Euler equation, continuity equation and the equation of state and linearise them. As an example we study several different velocity profiles: the so-called Rankine vortex model; the Burgers-Rott vortex model; Sullivan and modified Sullivan vortex models. It has been shown that in the two dimensional Rankine vortex model no instability appears in the inner region of a tornado. On the contrary, outside this area the physical system undergoes strong exponential instability. We have found that initially perturbed velocity components lead to amplified sound wave excitations. The similar results have been shown in Burgers-Rott vortex model as well. As it was numerically estimated, in this case, the unstable wave increases its energy by a factor of $400$ only in $\sim 0.5$min. According to the numerical study, in Sullivan and modified Sullivan models, the instability does not differ much by the growth. Despite the fact that in the inner area the exponential instability does not appear in a purely two dimensional case, we have found that in the modified Su
Using the Bernoulli integral for air streamline with condensing water vapor a stationary axisymmetric tornado circulation is described. The obtained profiles of vertical, radial and tangential velocities are in agreement with observations for the Mulhall tornado, world's largest on record and longest-lived among the three tornadoes for which 3D velocity data are available. Maximum possible vortex velocities are estimated.
(Abridge) On 13 September 2025 around 22 UTC, a localized tornado outbreak affected eastern Philippines, causing significant damage in Camarines Norte. The event developed within an atypical easterly severe weather regime characterized by warm, moist southeasterly flow and strong low-level wind shear associated with an easterly wave trough. A vorticity convergence zone along the inverted trough enhanced low-level rotation, while highly curved streamwise hodographs indicated a favorable environment for supercells and tornadogenesis. At least five vortices were identified, including three tornadic supercells. The Magang tornado was rated IF2.5 (EF3-equivalent) with $\sim$2 km damage path, while the Cahabaan and Napilihan tornadoes were rated IF1 (EF1-equivalent), with Cahabaan producing $\sim$3 km damage path. The remaining vortices were rated IF0 (EF0-equivalent). These tornadoes occurred simultaneously, indicating multiple discrete supercells within the same mesoscale environment and possible inflow-outflow interactions. Dual-polarization radar observations revealed Z$_\text{DR}$ and K$_\text{DP}$ columns, a debris signature in the Magang tornado, and a bounded weak echo region (BW
Rapid and accurate building damage assessment in the immediate aftermath of tornadoes is critical for coordinating life-saving search and rescue operations, optimizing emergency resource allocation, and accelerating community recovery. However, current automated methods struggle with the unique visual complexity of tornado-induced wreckage, primarily due to severe domain shift from standard pre-training datasets and extreme class imbalance in real-world disaster data. To address these challenges, we introduce a systematic experimental framework evaluating 79 open-source deep learning models, encompassing both Convolutional Neural Networks (CNNs) and Vision Transformers, across over 2,300 controlled experiments on our newly curated Quad-State Tornado Damage (QSTD) benchmark dataset. Our findings reveal that achieving operational-grade performance hinges on a complex interaction between architecture and optimization, rather than architectural selection alone. Most strikingly, we demonstrate that optimizer choice can be more consequential than architecture: switching from Adam to SGD provided dramatic F1 gains of +25 to +38 points for Vision Transformer and Swin Transformer families,
In 1999 the NWS began using the phrase "tornado emergency" to denote tornado warnings for storms with the potential to cause rare, catastrophic damage. After years of informal usage, tornado emergencies were formally introduced to 46 weather forecasting offices in 2014 as part of the impact-based warning (IBW) program, with a nationwide rollout occurring over the following years. In concert with the new tiered warning approach, the Warning Decision Training Division (WDTD) also introduced suggested criteria for when forecasters should consider upgrading a tornado warning to a tornado emergency, which includes thresholds of rotational velocity (VROT) and significant tornado parameter (STP). Although significant research has studied both tornado forecasting and tornado warning dissemination in the decade since, relatively little work has examined the effectiveness of the tornado emergency specifically. Our analysis of all 89 IBW tornado emergencies issued from 2014-2023 found that forecasters do not appear to follow the suggested criteria for issuance in the majority of cases, with only a handful of tornado emergencies meeting both the VROT and STP thresholds. Regardless, 70% of torn
Weather radar is the primary tool used by forecasters to detect and warn for tornadoes in near-real time. In order to assist forecasters in warning the public, several algorithms have been developed to automatically detect tornadic signatures in weather radar observations. Recently, Machine Learning (ML) algorithms, which learn directly from large amounts of labeled data, have been shown to be highly effective for this purpose. Since tornadoes are extremely rare events within the corpus of all available radar observations, the selection and design of training datasets for ML applications is critical for the performance, robustness, and ultimate acceptance of ML algorithms. This study introduces a new benchmark dataset, TorNet to support development of ML algorithms in tornado detection and prediction. TorNet contains full-resolution, polarimetric, Level-II WSR-88D data sampled from 10 years of reported storm events. A number of ML baselines for tornado detection are developed and compared, including a novel deep learning (DL) architecture capable of processing raw radar imagery without the need for manual feature extraction required for existing ML algorithms. Despite not benefitin
Magnetic tornadoes, characterized as impulsive Alfven waves initiated by photospheric vortices in intergranular lanes, are considered efficient energy channels to the corona. Despite their acknowledged importance for solar coronal heating, their observational counterparts from the corona have not been well understood. To address this issue, we use a radiative MHD simulation of a coronal loop with footpoints rooted in the upper convection zone, and synthesize the chromospheric and coronal emissions corresponding to a magnetic tornado. Considering SDO/AIA 171 A and Solar Orbiter/EUI 174 A channels, our synthesis reveals that the coronal response to magnetic tornadoes can be observed as an EUV brightening of which width is ~2 Mm. This brightening is located above the synthesized chromospheric swirl observed in Ca II 8542 A, Ca II K, and Mg II k lines, which can be detected by instruments such as SST/CRISP, GST/FISS, and IRIS. Considering the height correspondence of the synthesized brightening, magnetic tornadoes can be an alternative mechanism for the small-scale EUV brightenings such as the solar "campfires''. Our findings indicate that coordinated observations encompassing the chro
Tornadoes are among the most intense atmospheric vortex phenomena and pose significant challenges for detection and forecasting. Conventional methods, which heavily depend on ground-based observations and radar data, are limited by issues such as decreased accuracy over greater distances and a high rate of false positives. To address these challenges, this study utilizes the Seamless Hybrid Scan Reflectivity (SHSR) dataset from the Multi-Radar Multi-Sensor (MRMS) system, which integrates data from multiple radar sources to enhance accuracy. A novel hybrid model, the Kalman-Convolutional BiLSTM with Multi-Head Attention, is introduced to improve dynamic state estimation and capture both spatial and temporal dependencies within the data. This model demonstrates superior performance in precision, recall, F1-Score, and accuracy compared to methods such as K-Nearest Neighbors (KNN) and LightGBM. The results highlight the considerable potential of advanced machine learning techniques to improve tornado prediction and reduce false alarm rates. Future research will focus on expanding datasets, exploring innovative model architectures, and incorporating large language models (LLMs) to provi